Búsqueda Imágenes Maps Play YouTube Noticias Gmail Drive Más »
Iniciar sesión
Usuarios de lectores de pantalla: deben hacer clic en este enlace para utilizar el modo de accesibilidad. Este modo tiene las mismas funciones esenciales pero funciona mejor con el lector.

Patentes

  1. Búsqueda avanzada de patentes
Número de publicaciónUS5201318 A
Tipo de publicaciónConcesión
Número de solicitudUS 07/772,620
Fecha de publicación13 Abr 1993
Fecha de presentación4 Oct 1991
Fecha de prioridad24 Abr 1989
TarifaPagadas
También publicado comoUS5345941
Número de publicación07772620, 772620, US 5201318 A, US 5201318A, US-A-5201318, US5201318 A, US5201318A
InventoresJoseph J. Baraga, Michael S. Feld, Richard P. Rava, Rebecca Richards-Kortum
Cesionario originalRava Richard P, Richards Kortum Rebecca, Feld Michael S, Baraga Joseph J
Exportar citaBiBTeX, EndNote, RefMan
Enlaces externos: USPTO, Cesión de USPTO, Espacenet
Contour mapping of spectral diagnostics
US 5201318 A
Resumen
The present invention relates to a method of generating and processing spectral information arising from the laser induced fluorescence of tissue. The intensity of fluorescence is recorded as a function of both excitation and emission wavelengths and contour maps of tissue fluorescence are generated which are useful in the diagnosis of condition of the tissue under examination.
Imágenes(27)
Previous page
Next page
Reclamaciones(20)
We claim:
1. A method of processing spectra generated from the induced fluorescence of tissue comprising:
irradiating a portion of tissue over a range of periodically separated excitation wavelengths with radiation from a light source to induce autofluorescence of the tissue such that the tissue emits fluorescent radiation at a multiplicity of emission wavelengths;
forming a spectrum of the intensity of the emitted radiation from the recorded fluorescing radiation at each of the multiplicity of excitation and emission wavelengths;
repeating the irradiating and forming steps to provide a plurality of spectra;
averaging the plurality of spectra to produce an average spectrum; and
identifying fluorophores present in the tissue that have predetermined fluorescence characteristics which correlate with spectral features in the average spectrum, the feature having identified excitation and emission wavelengths present in the average spectrum.
2. The method of processing spectra of claim 1 wherein the tissue comprises arterial tissue.
3. The method of processing spectra of claim 1 wherein the tissue comprises colon tissue.
4. The method of processing spectra of claim 1 wherein the tissue comprises bladder tissue.
5. The method of processing spectra of claim 1 further comprising the step of forming a difference spectra by taking the difference between the formed spectrum and a reference spectrum.
6. The method of processing spectra of claim 1 further comprising identifying a region within the spectrum wherein the intensity in the region is reduced by a component of tissue which absorbs radiation.
7. The method of processing spectra of claim 6 further comprising adjusting the intensity in the region to correct for a contribution to the spectrum by the absorbing component of tissue.
8. The method of processing spectra of claim 1 further comprising the step of forming a ratio spectrum by dividing the formed spectrum by a reference spectrum.
9. A method of diagnosing bodily tissue comprising:
inserting a catheter containing an optical fiber into a body lumen and positioning a distal end of the catheter adjacent to tissue to be diagnosed;
coupling a proximal end of the catheter to a radiation source such that the radiation is transmitted along the optical fiber and directed onto the tissue to be diagnosed;
inducting autofluorescence with the tissue with the radiation source with a multiplicity of periodically separated excitation wavelengths;
recording the autofluorescent radiation emitted by the tissue at a multiplicity of emission wavelengths;
forming a spectral matrix of the intensity of the emitted radiation from the recorded radiation at each of the multiplicity of excitation and emission wavelengths;
repeating the inducing and recording steps to provide a plurality of the spectral matrices;
averaging the plurality of spectral matrix to produce an average spectral matrices; and
identifying fluorophores present in the tissue from the average spectral matrix to diagnose a condition of the tissue.
10. The method of claim 9 further comprising comparing the spectral matrix with a reference matrix to determine the presence of abnormal tissue.
11. The method of claim 9 wherein the radiation source comprises a laser.
12. The method of claim 9 wherein the lumen comprises an artery such that fluorophores of atherosclerotic material within the artery are identified.
13. The method of claim 9 further comprising identifying a portion of the spectral matrix wherein the intensity of the emitted radiation is reduced by a component of tissue which absorbs radiation.
14. The method of claim 13 further comprising adjusting the intensity of the identified portion of the spectral matrix to correct for the reduced intensity.
15. A method of diagnosing vascular tissue comprising:
inserting a catheter containing an optical fiber into a vascular lumen and positioning a distal end of the catheter adjacent to vascular tissue to be diagnosed;
coupling a proximal end of the catheter to a laser radiation source such that the laser radiation is transmitted along the optical fiber and directed onto the tissue to be diagnosed;
inducing inelastically scattered light to be emitted from the vascular tissue with the laser radiation with a multiplicity of periodically separated excitation wavelengths;
recording the scattered radiation returning from the tissue at a multiplicity of emission wavelengths;
forming a spectral matrix of the intensity of the emitted radiation from the recorded radiation at each of the multiplicity of excitation and emission wavelengths;
repeating the inducing and recording steps to provide a plurality of spectral matrices;
averaging the spectral matrices to provide an average spectral matrix; and
identifying abnormal tissue present in the tissue from the average spectral matrix.
16. The method of claim 15 further comprising comparing the spectral matrix with a reference matrix to determine the presence of abnormal tissue.
17. The method of claim 15 wherein the lumen comprises an artery such that fluorophores of atherosclerotic material within the artery are identified.
18. The method of claim 15 further comprising identifying a portion of the spectral matrix wherein the intensity of the emitted radiation is reduced by a component of tissue which absorbs radiation.
19. The method of claim 18 further comprising adjusting the intensity of the identified portion of the spectral matrix to correct for the reduced intensity.
20. The method of claim 15 wherein the inelastically scattered light comprises fluorescence.
Descripción
DETAILED DESCRIPTION OF THE INVENTION

FIG. 1 illustrates a preferred embodiment of the system that is used in generating and processing matrices of fluorescence intensity obtained as a function of both excitation and emission wavelengths. Light source 10 can be a broadband source such as a Xenon arc lamp. However, a tunable laser can also be employed to provide excitation radiation of variable wavelength. A double monochromator 12 subject to scan control 14 directs radiation of selected wavelength through optical fiber 16 onto the tissue sample 22. Collection fibers 18 collect the fluorescent radiation emitted by the sample 22 and direct it through collection optics 20 and into a second double monochromator 24. Photomultiplier 26 then detects the radiation and generates a signal that is forwarded to computer 28 for processing and display. Computer 28, via line 30, can be used to control the scanning parameters of the two monochromators 12 and 24.

The optical probe or catheter 32 can also be used in vivo within the human body to perform diagnostic scans on living tissue. The methods of the present invention may be used in conjunction with the laser catheter systems disclosed in U.S. Pat. No. 4,718,417, incorporated herein by reference.

Fluorescence intensity from normal aorta was recorded over a range of from 250 to 700 nm every 5 nm for both excitation and emission. The system used for this type of measurement is easily tuned over a large range of excitations. The system can use a rhodamine solution as a reference, so that differences in intensity of the exciting light at different wavelengths are automatically removed from the spectra. This makes it possible to obtain fluorescence intensities over a large range of excitations and emissions on a single sample of tissue in a reasonable amount of time. Gratings blazed at 250 nm have been used so that emission spectra can be taken from 250 to 700 nm.

To generate the data, emission scans were taken with excitations of 250 to 500 nm every 5 nm. The emission ranges ran from the excitation +12.5 nm to 697.5 nm every 5 nm. The FIGURES described below are generated by sequential acquisition of emission scans; however, excitation scans could be used as well to display the relationship between excitation and emission. The FWHM (full width-half maximum) of the excitation source was 2 nm, and the emission FWHM was 3 nm. An integration time of 1 second at each point yielded a reasonable signal to noise ratio for all scans.

FIG. 2 shows the contour map that was generated from a sample of morphologically normal aorta. The map shows five distinct peaks (labeled A-E in FIG. 2). Each peak potentially represents a distinct fluorophore; however, it is known that fluorescence reabsorption in the tissue can create artificial peaks in emission spectra. On the contour map, such reabsorption valleys should appear as depressions at constant emission wavelength, since the wavelength of reabsorption does not depend on excitation. The three dashed vertical lines in FIG. 2 are examples of such reabsorption. The two lines near 570 nm emission represent absorption due to oxy-hemoglobin, while the line near 415 nm is the Soret band of heme. Based on this interpretation, the two peaks D and E are due to a single fluorophore which is known to be elastin/collagen. The peaks B and C may also be due to a single fluorophore, although the fact that they are slightly shifted in excitation from one another supports the notion that the peaks are due to two distinct fluorophores.

Horizontal lines, that is, lines of constant excitation, on the map represent emission scans. Four emission scans from the region designated by line A-B in FIG. 2 are shown in FIG. 3a. These scans are very similar to those with 476 nm excitation. The two peaks at 560 and 600 nm are artificial in that they are created by the reabsorption valleys at 540 and 580 nm, as was surmised from the contour map.

Emission scans from line A-C are shown in FIG. 3b. As can be seen from the contour map, the intensity of fluorophore A, emitting at 340 nm, is strongly dependent on excitation wavelength near 300 nm. This is vident from the very close spacing of the contour lines in this region. This fluorophore has been identified as tryptophan. The Soret valley at 415 nm is also seen in these emission scans, as is the emergence of another peak (B) at 390 nm, as excitation wavelength is made longer. Based on spectra from thin samples of artery, peak C (at 450 nm emission) was assumed to be artificially created by the Soret valley, so that peaks B and C were assigned to a single fluorophore. However, as the contour map reveals, these peaks occur at slightly different excitations, and it is possible that two distinct fluorophores are being observed.

This system has been used to determine optimal excitation and emission wavelengths for differentiating normal and atherosclerotic aorta. Also, contour maps of fibrous, fatty and calcified plaques have been constructed. By normalizing these maps and then subtracting them from, or ratioing them, to a normalized map of normal aorta, excitation and emission wavelengths at which fluorescence lineshapes of normal and diseased tissue are most different can be easily identified.

FIGS. 4-7 present data collected as excitation-emission matrices (EEM's) of bulk samples of normal and pathological human aorta tissue. This data, which is presented in the form of contour maps as well as perspective drawings is used to determine an optimum wavelength for distinguishing between normal and pathological tissue.

Tissue samples are classified as either normal, fibrous, fatty, or calcified. In all, there were nine normal samples run from five different patients six calcified samples which were all from the same patient, eight fibrous samples taken from four different patients, and two samples of fatty tissue from the same patient.

The EEM's of each tissue type were averaged with the exception of the fatty sample shown in FIG. 7

The main features of the EEM's of the normal samples of FIGS. 4a and 4b are the dominant tryptophan peak in the uv and the absorption valleys at 415 nm, which are believed to be caused by the presence of hemoglobin in the tissue. There is also weaker fluorescence in the visible range, which is primarily due to chromophores within the collagen and elastin. The most noticeable difference between the EEM's of the normal and fibrous tissue (at FIGS. 5a and 5b) is that there is greater fluorescence at the longer wavelengths in normal compared to fibrous. There also appears to be a difference in the location of the peak of the fluorescence due to chromophores within the collagen, but this may be the result of less hemoglobin absorption in the fibrous tissue. One difference between the calcified tissue shown in FIGS. 6a and 6b and the normal tissue is that the fluorescence peak due to tryptophan is about a factor of two lower in the calcified tissue. In addition, the fluorescence due to chromophores within the collagen and elastin is two to three times greater in the calcified tissue, and there is also less absorption at 415 nm. One of the chromophores found within the collagen or elastin morphologic structures is pyridinoline which is known to fluoresce in the range 400-420 nm when excited by radiation in the range of 360-370 nm.

Hemoglobin (Hb) plays an important role in in vitro aorta EEM's. The large Soret absorption band at 416 nm is responsible for the deep valleys seen in the contour map both at excitations and emissions around 416 nm. Oxyhemoglobin also produces the minor valleys near 540 and 580 nm. Note that heme is one of the chromophores found in hemoglobin. Heme has an absorption region centered around 560 nm. This conclusion is supported by the following observations:

1) The absorption spectrum of hemoglobin has peaks coincident with the valleys mentioned above (FIGS. 8a and 8b).

2) In vitro aorta samples have typically been exposed to lysed red blood cells and, therefore, free Hb which can readily diffuse into the tissue.

3) These valleys are absent in 10 μm sections of intima and media. The effects of absorption on LIF are negligible in these thin sections.

4) These valleys can be nearly eliminated by soaking an in vitro bulk aorta sample in saline before collecting the spectra (FIGS. 9a and 9b). This soaking removes most of the Hb, which can freely diffuse from the tissue.

The physical basis for these valleys is quite simple: a portion of the incident excitation light as well as the emitted fluorescence is reabsorbed by the Hb, causing a reduction in the fluorescence intensity. This can be clearly seen from the single layer tissue fluorescence model, in which the observed fluorescence intensity S(λ.sub.x,λm) can be written ##EQU1## where F(λ.sub.x,λm) is the "intrinsic" tissue fluorescence, i.e. the tissue fluorescence from a thin section, free of absorption effects; x.sub.Hb A.sub.Hb (λ.sub.x,λm)=μ.sub.Hb (λ.sub.x) +μ.sub.Hb (λ.sub.m), where A.sub.Hb represents the attenuation spectrum of Hb (and similarly for SP, where SP refers to fixed tissue absorbers). A.sub.Hb (λ.sub.x,λm) can be thought of as the normalized attenuation matrix of Hb, with x.sub.Hb representing its concentration; it can be generated from the Hb absorption spectrum if effects of scattering are neglected. A contour map of A.sub.Hb (λ.sub.x,λm), which clearly shows the Soret absorption bands, is shown in FIG. 10.

Equation (1) can be rewritten ##EQU2## where S.sub.0 (λ.sub.x,λm) is the fluorescence observed in the absence of Hb(x.sub.Hb= 0). By making the crude assumption that A.sub.SP is a constant (not dependent on wavelength) over the entire EEM, by using values of x.sub.Hb and x.sub.SP A.sub.SP determined with 476 nm excitation, and by using the calculated Hb absorption matrix, the effect of Hb absorption can be reproduced from the soaked aorta EEM (x.sub.Hb= 0) by "dividing" Hb absorption matrix into it in the manner specified by Eq. 2. The result is displayed in FIGS. 11a and 11b. The structure of the aorta EEM with Hb (before soaking) is recovered, although peak intensities, especially in the ultraviolet, are smaller than in the actual EEM.

Using this method, Hb absorption effects can also be produced in an ex vivo aorta sample. The EEM of the ex vivo sample which has presumably not been exposed to free Hb, is quite similar to that of the soaked aorta. By dividing the Hb absorption matrix into it in the manner of Eq. 2, an EEM very similar to that of Hb-containing in vitro aorta is produced.

This method of correcting for the presence of absorbers within the tissue being examined can be applied to other absorbing components. The removal of the effects of absorption can assist in providing a clearer spectrum in which all of the avisibile information regarding tissue fluorescence can be recovered.

The system also assists in identifying the chromophores which contribute to arterial tissue fluorescence spectra. Contour maps provide very useful insight about the number and nature of tissue chromophores. To separate contributions from different chromophores, maps can be obtained from thin tissue sections, which contain a limited number of chromophores (e.g. ceroid in necrotic core) and are free of reabsorption effects. Also maps of pure compounds, which are suspected tissue chromophores, can be constructed for comparison with tissue maps. For example, FIG. 12 contains a map of powdered type 1 bovine collagen. Other components studied include chromophores or structures containing chromophores such as elastin, tryptophan, hemoglobin and flavoprotein.

Although type 1 collagen has emission in the same region as does normal aorta, there are several important differences. Some of these are due to reabsorption of hemoglobin; the collagen map does not show valleys at 420, 540 and 580 nm. However, other differences are not related to reabsorption, and indicate that tissue fluorescence in this region contains contributions from chromophores other than those present in type 1 collagen.

The methods of the present invention have been used to differentiate normal and pathologic tissues in human colon and urinary bladder with excitation wavelengths ranging from 250-500 nm, utilizing fluorescence contour maps. This contour mapping serves to identify optimal excitation wavelengths for differentiating normal and pathologic tissue fluorescence spectra and to identify tissue chromophores contributing to the fluorescence of normal and pathologic tissues.

Fluorescence contour maps were collected from 18 colon specimens and 15 urinary bladder specimens. Fluorescence contour maps were constructed from a series of fluorescence emission spectra recorded using a spectrofluorimeter. Excitation wavelengths were varied in 10 nm steps from 250 to 500 nm. Fluorescence was collected at 5 nm intervals from λexc+10 nm to 2λexc-10 nm. Incident beam size was ˜2 thus absolute intensity information has been preserved in the analysis of this data. To correct for day to day variations in the spectrofluorimeter alignment, all data has been divided by a fluorescence intensity standard which is run each day. Emission gratings were blazed at 250 nm. All data presented here has been corrected for the non-uniform spectral response of the collection system.

The following method of data analysis was employed for both colon and bladder tissues. An average normal contour map N.sub.A (λ.sub.x,λm), and an average pathologic contour map P.sub.A (λ.sub.x,λm) were calculated and were plotted on a log scale with contours ranging from 50 to 0.5. To compare contour maps of normal and pathologic tissues, a ratio of the average contour maps was constructed as: ##EQU3## and was plotted on a linear scale with contours varying from 2.0 to 0.2. Here, a contour at 1.0 indicates regions in which the fluorescence emission from normal and pathologic tissues is the same, while contours at values greater (less) than 1.0 indicate regions where the emission from pathologic tissue is greater (less) than that from normal tissues Regions of greatest difference in the fluorescence spectra of normal and pathologic tissues were assessed from this average ratio map. This information was used in defining a process for determination of tissue type.

To assess how the variability of individual tissues affected this comparison, the following ratio maps were constructed for each individual normal and pathologic sample: ##EQU4## Data from these individual ratio maps was used to test processes suggested in the above data analysis procedure.

FIGS. 13 and 14 show average fluorescence contour maps of normal and adenomatous human colon tissue. In the normal tissue fluorescence contour map, several major fluorescence and absorption bands can be recognized, as summarized in Table 1. Similarly, Table 2 lists the major absorption and fluorescence bands found in the adenomatous fluorescence contour map.

              TABLE 1______________________________________Average Normal Colon Tissue       Fluorescence(λx, λm)       Intensity    Chromophore______________________________________(285, 340)  50.0         Tryptophan(330, 385)  5.0          Chromophore                    Within Collagen(315, 430)  2.5          4-Pyridoxic Acid(345, 465)  5.0          NAD(P)H(460, 520)  4.5          Flavin(λ, 420) (420, λ)                    Hemoglobin (Heme)(λ, 540) (λ, 580)______________________________________

              TABLE 2______________________________________Average Adenomatous Colon Tissue       Fluorescence(λx, λm)       Intensity    Chromophore______________________________________(285, 340)  50.0         Tryptophan(340, 470)  2.5          NAD(P)H(460, 520)  2.5          Flavin(λ,420) (420, λ)                    Hemoglobin (Heme)(λ, 540) (λ, 580)______________________________________

A comparison of these two tables points out several of the interesting differences in the fluorescence contour maps of normal and adenomatous tissues Both tissues exhibit tryptophan fluorescence. Qualitatively, the shape and peak fluorescence intensity of this band is very similar for both types of tissue Two additional fluorescence bands appear to be present in the normal tissue fluorescence contour map at (330,385) and (315,430). In the visible region of the spectrum, the fluorescence intensity of normal tissues is approximately 2 tissues. In particular, both tissues show a band near (340,470); this band is 2 slightly shifted in normal tissue relative to adenomatous tissue. Both tissues show a peak at (460,520), which is 1.8.times. more intense in normal tissue.

These differences are highlighted in FIG. 15, which shows the ratio of the two average contour maps, with contours drawn linearly from 2.0 to 0.2. Greater detail is shown in FIG. 16, with contours drawn from 1.0 to 0.1. Again contours at 1.0 indicate regions where normal and adenomatous tissues exhibit equal fluorescence intensities, while contours greater (less) than 1.0 indicate regions where adenomatous tissue fluorescence is greater (less) than normal tissue fluorescence. In the region of tryptophan fluorescence, a contour with a value of 1.0 is shown, indicating the similarity of tryptophan fluorescence in normal and adenomatous tissues. A valley is present at (335,385) with a contour level of 0.3 indicating the additional fluorescence band present in normal tissue at this location. A valley is present at (390,430) with a contour of 0.8. Additional valleys are present at (400,495), (440,480), (430,600) and (430,670). These are summarized in Table 3.

              TABLE 3______________________________________Average Ratio Map, Colon Tissue(λx, λm)     .sup.R AVG.sup.(λ x, .sup.λ m.sup.)                    Chromophore______________________________________(330, 385)     0.3            Chromophore                    Within Collagen(390, 430)     0.8            Pyridoxic Acid                    Lactone(400, 495)     0.4            ?(440, 480)     0.4            ?(430, 600)     0.7            ?(430, 670)     0.8            ?______________________________________

The values of the individual ratio maps at these locations of (λ.sub.x,λm), can be divided by the intensity of the ratio map at (290,340). Two of these values, at (330,385) and (440,480), independently provided useful diagnostics for differentiating normal and adenomatous tissues (FIGS. 17 and 18). In addition, a binary scheme utilizing information at (390,430) and (330,385) provided an accurate diagnostic process (FIG. 19).

FIGS. 20 and 21 show average fluorescence contour maps of normal bladder wall and tumor. Major fluorescence and absorption bands are summaried in Tables 4 and 5. Again, in the region of tryptophan fluorescence both normal and tumor tissues have similar emission. Although both tissues show a fluorescence band at (325,385), it is 3 tissue. Both tissues show fluorescence peaks at (350,470) with normal tissues exhibiting 2 (470,520) the fluorescence of normal tissue is 2 tumor tissue. The average tumor map exhibits a unique peak at (315,430).

              TABLE 4______________________________________Average Normal Bladder Tissue       Fluorescence(λx, λm)       Intensity    Chromophore______________________________________(285, 340)  50.0         Tryptophan(325, 385)  15.0         Chromophore                    Within Collagen(350, 470)  10.0         NAD(P)H(470, 520)  10.0         Flavin(λ, 420) (420, λ)                    Hemoglobin (Heme)(λ, 540) (λ, 580)______________________________________

              TABLE 5______________________________________Average Bladder Tumor       Fluorescence(λx, λm)       Intensity    Chromophore______________________________________(285, 340)  50.0         Tryptophan(330, 385)  5.0          Chromophore                    Within Collagen(315, 430)  3.0          4-Pyridoxic Acid(350, 470)  5.0          NAD(P)H(460, 515)  5.0          Flavin(λ, 420) (420, λ)                    Hemoglobin (Heme)(λ, 540) (λ, 580)______________________________________

These differences are highlighted in FIGS. 22 and 23 which show the ratio contour map of tumor and normal bladder tissue for two sets of contours (2.0-0.2, 1.0-0.1). Table 6 summarizes the valleys present in this ratio map. Valleys are present at (330,385), (370,435), (400,495), (450,485), (420,640), and (390,670). Values of the individual ratio maps at these locations can be divided by the value of the tryptophan peak (290,340). FIG. 24 shows an example of a diagnostic procedure which can be defined from this data.

              TABLE 6______________________________________Average Ratio Map, Bladder Tissue(λx, λm)     .sup.R AVG.sup.(λ x, .sup.λ m.sup.)                    Chromophore______________________________________(330, 385)     0.4            Chromophore                    Within Collagen(370, 435)     0.8            Pyridoxic Acid                    Lactone(400, 495)     0.5            ?(450, 485)     0.4            ?(420, 640)     0.9            ?(390, 670)     0.6            ?______________________________________

Because of the voluminous information in the 2-D plots, a model is useful in interpreting the data. Even if the types and concentrations of all chromophores were known, quenching, spectral shifts, and energy transfer due to the local environment would alter a calculated spectrum. But, in spite of these many unknowns, a "landscape" model of chromophore behavior does provide some general insights for locating distinctive features for comparative spectral diagnosis.

The model begins with the reasonable assumption that tissue contains a few relatively strongly fluorescing chromophores, a background of weaker ones, and absorbers which are probably non-fluorescent. The strong chromophores, with relatively high quantum yields or concentrations (or both), will generate "peaks" on the contour plot. In addition, these chromophores generally have a broad base which tends toward shorter wavelengths for excitation; and due to the general anharmonicity of higher vibrational levels, it tends toward longer wavelengths for emission. Combining with the weaker chromophores, this forms an average elevation of signal background from which the peaks arise. The absorbers cut vertically and horizontally across the spectral terrain. The percent decrease along the line should be approximately constant to the extent that the penetration depth is constant. This is distinguishable from a "pass" between two spectral peaks, which is just lesser fluorescence, not absorption. Thin sections will be quite helpful in identifying absorption features by their absence.

Distinguishing diseased vs. healthy tissue is essentially distinguishing one set of chromophore signals from another. If two tissues have different amounts of one chromophore, then the corresponding absolute signals might be measured; however, this is dependent on collection geometry. Comparisons, or ratios, are less geometry dependent, but appropriate positions on the 2-D surface need to be selected. The data is normalized so as to again avoid geometry-dependent ratios.

The spectral landscape model has three regions: 1) fluorescent peaks, 2) an average elevation, or plateau, and 3) absorption valleys. Of these, region 1 is the most appealing as a discriminant for tissue type, as it would show a characteristic chromophore. Region 3 is susceptible to variability or leaching as has been shown for the case of heme, and region 2, being comprised of fluorescence from several chromophores, is likely to have the least variation overall. Signal change due to a change in one of the chromophores would be diminished by the fraction of the total signal contributed by that chromophore. In other words, instead of comparing absolute peak heights, they should first be normalized relative to the average elevation. Ideally, normalization would be relative to selected regions thought to be "average"; a practical compromise is the average height from the total integrated signal of the entire plot. Peaks then have signals >1, and valleys generally <1. The comparison signal Rλ for each set of excitation and emission wavelengths, equals the diseased signal Dλ, divided by the normal tissue signal, Nλ.

Rλ=Dλ/Nλ

If subtraction is used instead of the ratio, then this number should be normalized to the signal strength at that wavelength to indicate the fractional, rather than absolute, change:

Sλ=ΔDλ/Nλ,

where ΔDλ=Dλ-Nλ and therefore Sλ=Rλ-1.

This gives the normalized differential signal Sλ. N is in the denominator, since it is desirable to determine the deviation of the diseased tissue from normal. When the "peak" value is selected, it should be integrated over the top 10%-20% of the peak, thereby incorporating a substantial volume of the fluorescence signal; otherwise, data is discarded, and the noise level is increased.

An alternative approach is to normalize two contour plots to the same chromophore peak, then ratio them. This would compare amplitudes of other peaks to the first. However, the ratioed contour plot is dependent on which peak is selected for normalization, and the information from the rest of the plot, which provides the total integrated signal, is in effect, unused.

Additional information is contained in the peak shape. If the different environments of the tissue cause spectral shifts, then this may be emphasized by ratioing normalized peaks on two contour plots. In this case a smaller integrated area may be better for normalization. Deviations from unity will indicate different slopes or shapes.

To avoid overemphasizing weaker parts of the spectrum, a weighting factor may be introduced: Assuming that the signal to noise ratio varies as the square root of the signal, the ratio signal should be multiplied by the square root of Nλ to yield Wλ, the weighted normalized ratio contour plot comparing the two tissue types.

Wλ=NλRλ.

These computations should be straightforward for numerical peak comparison. Large differences in regions where the fluorescence is weak will be de-emphasized.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of an apparatus used in conjunction with the present invention.

FIG. 2 illustrates a contour map of morphologically normal aorta.

FIG. 3a presents four emission scans from the region about line A-B of FIG. 2.

FIG. 3b presents emission scans from line A-C in FIG. 2.

FIGS. 4a and 4b present a contour map and a perspective view of an excitation emission matrix of fluorescence intensity of normal aorta.

FIGS. 5a and 5b show a contour map and a perspective view, respectively, of an averaged excitation-emission matrix of fibrous aortic tissue.

FIGS. 6a and 6b show a contour map and a perspective view, respectively, of an averaged excitation-emission matrix of calcified aortic tissue.

FIGS. 7a and 7b show a contour map and a perspective view, respectively, of an excitation-emission matrix of fatty aortic tissue.

FIGS. 8a and 8b show an absorption spectrum and a contour map, respectively, of hemoglobin.

FIGS. 9a and 9b show contour maps of bulk aorta before and after the sample is soaked in saline.

FIG. 10 shows a contour map of the attenuation matrix of hemoglobin.

FIGS. 11a and 11b show contour maps of aortic tissue after soaking to remove hemoglobin, and after the addition of the effects of hemoglobin using the hemoglobin absorption matrix.

FIG. 12 illustrates a contour 12 of powdered type 1 bovine collagen.

FIGS. 13a and 13b show average fluorescence contour maps of normal human colon tissue.

FIGS. 14a and 14b show average fluorescence contour maps for adenomatous human colon tissue.

FIG. 15 shows the ratio of an average contour map for adenomatous polyp to an average contour map for normal gastrointestinal tissue.

FIG. 16 shows the map of FIG. 15 where the contour line density is increased by a factor of two to show more detail.

FIG. 17 illustrates a diagnostic algorithm for differentiating adenomatous from normal colon tissue based upon a ratio of maps at an excitation wavelength of 330 nanometers and an emission wavelength of 385 nanometers.

FIG. 18 illustrates a diagnostic algorithm based upon an excitation wavelength of 440 nanometers and an emission wavelength of 480 nanometers.

FIG. 19 illustrates a binary type diagnostic algorithm.

FIGS. 20a and 20b show average contour maps on different scales of normal bladder tissue.

FIGS. 21a and 21b show average contour maps on two different scales of bladder tumor.

FIG. 22 shows a ratio of contour maps for bladder tumor and normal bladder tissue.

FIG. 23 shows the map of FIG. 22 with a greater contour line density.

FIG. 24 shows a binary diagnostic algorithm to differentiate bladder tumor from normal bladder tissue.

BACKGROUND

The present invention relates to a system and method for processing spectral information used to aid in the diagnosis of diseased tissue. More particularly, it relates to a method for collecting and displaying the excitation and emission spectra resulting from the laser-induced-fluorescence of tissue.

Methods utilizing the laser-induced-fluorescence ("LIF") of tissue have been developed which permit the characterization of the tissue being examined. This technique has traditionally employed the use of fluorescing agents or dyes which are introduced or applied to the tissue of interest which is then irradiated to induce fluorescence and produce a spectrum that can be used to distinguish diseased from normal tissue. The emission spectra obtained through these methods are normally plotted as intensity versus emission wavelength for a given excitation wavelength. These graphical displays can provide information regarding the diseased condition of tissue as specific peaks in the spectra of abnormal tissue appear which are not present in normal tissue.

Laser catheter systems have also been developed, often in conjunction with known techniques for using fluorescence to aid in diagnosis, for the purpose of inserting a light transmitting device into the human body where laser radiation can be directed onto tissue adjacent the distal end of the catheter to induce fluorescence. The light emitted by the tissue is transmitted along the catheter and analyzed at the proximal end thereof to produce an emission spectrum of the irradiated tissue.

The problem with these standard spectra is that they are highly dependent upon the excitation wavelength that is used such that a highly fragmented view of the spectral characteristics of the tissue is provided. Thus, a more complete method of generating diagnostic information utilizing LIF spectroscopy is needed that more fully characterizes the spectral features of tissue.

SUMMARY OF THE INVENTION

The present invention relates to a system for characterizing tissue fluorescence in the ultraviolet and throughout the visible for the purpose of generating diagnostically useful information. The present system provides the ability to look at fluorescence over a wide range of excitation and emission wavelengths. Such a broad survey is important in fully characterizing the spectroscopic differences between tissue types as well as in aiding in the identification of tissue fluorophores whose presence is correlated with known normal and abnormal states of tissue.

The intensity of fluorescence depends on three parameters: the radiative properties of the fluorophores within the tissue; the wavelength of the exciting light; and the wavelength of the emitted fluorescence. Two types of scans are typically used when recording fluorescence intensity: emission scans (spectra), in which the excitation wavelength is held constant and the emission wavelength is varied, and excitation scans (spectra), in which the emission wavelength is constant and the excitation is varied. Either type of scan alone contains useful information about the tissue constituents; however, both excitation and emission profiles are needed for a full characterization of fluorescence. In fact, a complete fluorescence characterization involves recording the fluorescence intensity for each possible pair of excitation and emission wavelengths.

Typically, emission (or excitation) scans are presented as two dimensional plots with emission (or excitation) wavelengths running along one axis and fluorescence intensity running along the other axis. In order to display intensity for each possible pair of excitation and emission wavelengths, a 3-dimensional plot of intensity vs. excitation and emission is one alternative for displaying the information for analysis. There are also two methods employed for representing such plots in two dimensions: perspective drawing and contour (topographic) mapping. In a preferred embodiment, contour maps are employed to display the data as they provide for viewing many of the important features on one plot.

This method is employed in conjunction with laser catheter systems used to induce fluorescence in bodily tissue both in vitro and in vivo without the use of fluorescence enhancing agents, also referred to herein as "autofluorescence". The procedure is illustrated using data collected on excised tissue samples, including artery wall, gastrointestinal tissue and bladder tissue, but are usefully applied to all types of tissue and the abnormal or diseased conditions which can be differentiated by optical characteristics.

In particular, fluorophores present in the tissue under study, whose concentration varies with the condition of the tissue, have been associated with specific features of the contour spectra. This correlation between the contour spectra and the chemical constituents of the tissue is useful in the construction of diagnostic methods which can systematically differentiate tissue type and condition.

The above and other features of the invention, including various novel details in the methods described and certain combinations thereof, will now be more particularly described with reference to the accompanying drawings and pointed out in the claims. It will be understood that the particular methods of diagnosis embodying the invention are shown by way of illustration only and not as a limitation of the invention. The principle features of the invention may be employed in various embodiments without departing from the scope of the invention.

This is a continuation of co-pending application Ser. No. 07/342,311 filed on Apr. 24, 1989, now abandoned.

Citas de patentes
Patente citada Fecha de presentación Fecha de publicación Solicitante Título
US4178917 *3 Ene 197918 Dic 1979Shapiro Howard MMethod and system for non-invasive detection of zinc protoporphyrin in erythrocytes
US4449535 *25 Mar 198222 May 1984Compagnie Industrielle Des Lasers Cilas AlcatelApparatus for measuring in situ the state of oxidation-reduction of a living organ
US4479499 *29 Ene 198230 Oct 1984Alfano Robert RMethod and apparatus for detecting the presence of caries in teeth using visible light
US4556057 *11 Mar 19833 Dic 1985Hamamatsu Tv Co., Ltd.Cancer diagnosis device utilizing laser beam pulses
US4648892 *22 Mar 198510 Mar 1987Massachusetts Institute Of TechnologyMethod for making optical shield for a laser catheter
US4768516 *18 Feb 19866 Sep 1988Somanetics CorporationMethod and apparatus for in vivo evaluation of tissue composition
US4785806 *8 Ene 198722 Nov 1988Yale UniversityLaser ablation process and apparatus
US4786813 *22 Oct 198522 Nov 1988Hightech Network Sci AbFluorescence imaging system
US4894547 *28 Sep 198716 Ene 1990Yale UniversityOptical method and apparatus for detecting and measuring aging, photoaging, dermal disease and pigmentation in skin
US4930516 *25 Abr 19885 Jun 1990Alfano Robert RMethod for detecting cancerous tissue using visible native luminescence
US4957114 *28 Jun 198818 Sep 1990Kun ZengDiagnostic apparatus for intrinsic fluorescence of malignant tumor
US5003977 *9 Mar 19892 Abr 1991Agency Of Industrial Science And TechnologyDevice for analyzing fluorescent light signals
US5042494 *4 Dic 198927 Ago 1991Alfano Robert RMethod and apparatus for detecting cancerous tissue using luminescence excitation spectra
US5046501 *18 Ene 198910 Sep 1991Wayne State UniversityAtherosclerotic identification
US5062431 *30 Mar 19905 Nov 1991Health Research, Inc.In vivo fluorescence photometer
Otras citas
Referencia
1 *Pottier et al., Photochemistry & Photobiology, vol. 44, No. 5, pp. 679 687 (1986).
2Pottier et al., Photochemistry & Photobiology, vol. 44, No. 5, pp. 679-687 (1986).
3Publication, "Tumour Localization by Means of Laser-Induced Fluorescence in Hematoporphyrin Derivative (HPD) Bearing Tissue".
4 *Publication, Tumour Localization by Means of Laser Induced Fluorescence in Hematoporphyrin Derivative (HPD) Bearing Tissue .
Citada por
Patente citante Fecha de presentación Fecha de publicación Solicitante Título
US5345941 *1 Mar 199313 Sep 1994Massachusetts Institute Of TechnologyContour mapping of spectral diagnostics
US5363855 *25 Ago 199215 Nov 1994Rutgers UniversityPressure waveform monitor
US5421337 *29 Mar 19946 Jun 1995Massachusetts Institute Of TechnologySpectral diagnosis of diseased tissue
US5421339 *12 May 19936 Jun 1995Board Of Regents, The University Of Texas SystemDiagnosis of dysplasia using laser induced fluoroescence
US5612540 *31 Mar 199518 Mar 1997Board Of Regents, The University Of Texas SystemsOptical method for the detection of cervical neoplasias using fluorescence spectroscopy
US5640957 *9 Jun 199524 Jun 1997Instruments Sa, Inc.Ultraviolet radiation protection evaluator
US5697373 *14 Mar 199516 Dic 1997Board Of Regents, The University Of Texas SystemOptical method and apparatus for the diagnosis of cervical precancers using raman and fluorescence spectroscopies
US5699795 *31 Mar 199523 Dic 1997Board Of Regents, The University Of Texas SystemOptical probe for the detection of cervical neoplasia using fluorescence spectroscopy and apparatus incorporating same
US5713364 *1 Ago 19953 Feb 1998Medispectra, Inc.Spectral volume microprobe analysis of materials
US5813987 *24 Dic 199629 Sep 1998Medispectra, Inc.Spectral volume microprobe for analysis of materials
US5842995 *28 Jun 19961 Dic 1998Board Of Regents, The Univerisity Of Texas SystemSpectroscopic probe for in vivo measurement of raman signals
US5984861 *29 Sep 199716 Nov 1999Boston Scientific CorporationEndofluorescence imaging module for an endoscope
US5991653 *19 Jun 199623 Nov 1999Board Of Regents, The University Of Texas SystemNear-infrared raman spectroscopy for in vitro and in vivo detection of cervical precancers
US6055451 *12 Dic 199725 Abr 2000Spectrx, Inc.Apparatus and method for determining tissue characteristics
US6095982 *11 Dic 19971 Ago 2000Board Of Regents, The University Of Texas SystemSpectroscopic method and apparatus for optically detecting abnormal mammalian epithelial tissue
US6096065 *29 Sep 19971 Ago 2000Boston Scientific CorporationSheath for tissue spectroscopy
US6104945 *13 Ene 199715 Ago 2000Medispectra, Inc.Spectral volume microprobe arrays
US6119031 *22 Jul 199712 Sep 2000Boston Scientific CorporationMiniature spectrometer
US6135965 *2 Dic 199624 Oct 2000Board Of Regents, The University Of Texas SystemSpectroscopic detection of cervical pre-cancer using radial basis function networks
US61742919 Mar 199816 Ene 2001Spectrascience, Inc.Optical biopsy system and methods for tissue diagnosis
US618544329 Sep 19976 Feb 2001Boston Scientific CorporationVisible display for an interventional device
US6241672 *18 Dic 19975 Jun 2001University Of WashingtonMethod and apparatus for optically imaging solid tumor tissue
US628922919 Ene 199911 Sep 2001Scimed Life Systems, Inc.Readable probe array for in vivo use
US632441829 Sep 199727 Nov 2001Boston Scientific CorporationPortable tissue spectroscopy apparatus and method
US63432276 Ene 200029 Ene 2002Boston Scientific CorporationMiniature spectrometer
US636483118 Ago 19992 Abr 2002Boston Scientific CorporationEndofluorescence imaging module for an endoscope
US63832091 Jun 20007 May 2002Boston Scientific CorporationSheath for tissue spectroscopy
US638548415 Dic 20007 May 2002Medispectra, Inc.Spectroscopic system employing a plurality of data types
US639331511 Sep 199821 May 2002Communaute EuropeenneDetecting and mapping of inflamed zones in a living tissue
US640507318 Oct 200011 Jun 2002Scimed Life Systems, Inc.Miniature spectrometer system and method
US641183822 Dic 199925 Jun 2002Medispectra, Inc.Systems and methods for optical examination of samples
US6671540 *3 Ago 199930 Dic 2003Daryl W. HochmanMethods and systems for detecting abnormal tissue using spectroscopic techniques
US669765219 Ene 200124 Feb 2004Massachusetts Institute Of TechnologyFluorescence, reflectance and light scattering spectroscopy for measuring tissue
US66976666 Mar 200024 Feb 2004Board Of Regents, The University Of Texas SystemApparatus for the characterization of tissue of epithelial lined viscus
US676061324 Jun 20026 Jul 2004Medispectra, Inc.Substantially monostatic, substantially confocal optical systems for examination of samples
US676891810 Jul 200227 Jul 2004Medispectra, Inc.Fluorescent fiberoptic probe for tissue health discrimination and method of use thereof
US681890313 Sep 200216 Nov 2004Medispectra, Inc.Method and apparatus for identifying spectral artifacts
US682642211 Ene 200030 Nov 2004Medispectra, Inc.Spectral volume microprobe arrays
US683966115 Dic 20004 Ene 2005Medispectra, Inc.System for normalizing spectra
US68474909 Jun 200025 Ene 2005Medispectra, Inc.Optical probe accessory device for use in vivo diagnostic procedures
US688287518 Oct 200019 Abr 2005Boston Scientific CorporationVisible display for an interventional device
US690293515 Dic 20007 Jun 2005Medispectra, Inc.Methods of monitoring effects of chemical agents on a sample
US691241218 Ene 200228 Jun 2005Massachusetts Institute Of TechnologySystem and methods of fluorescence, reflectance and light scattering spectroscopy for measuring tissue characteristics
US693315415 Nov 200223 Ago 2005Medispectra, Inc.Optimal windows for obtaining optical data for characterization of tissue samples
US710340110 Jul 20025 Sep 2006Medispectra, Inc.Colonic polyp discrimination by tissue fluorescence and fiberoptic probe
US712728212 Abr 200224 Oct 2006Medispectra, Inc.Optical methods and systems for rapid screening of the cervix
US713067225 Sep 200131 Oct 2006Critisense Ltd.Apparatus and method for monitoring tissue vitality parameters
US713651818 Abr 200314 Nov 2006Medispectra, Inc.Methods and apparatus for displaying diagnostic data
US718781018 Oct 20026 Mar 2007Medispectra, Inc.Methods and systems for correcting image misalignment
US726024815 Mar 200221 Ago 2007Medispectra, Inc.Image processing using measures of similarity
US728272318 Abr 200316 Oct 2007Medispectra, Inc.Methods and apparatus for processing spectral data for use in tissue characterization
US730228914 Jun 200127 Nov 2007Scimed Life Systems, Inc.Readable probe array for in-vivo use
US730777424 Ene 200611 Dic 2007The Board Of Trustees Of The Leland Standford Junior UniversityMicro-optical analysis system and approach therefor
US730986718 Abr 200318 Dic 2007Medispectra, Inc.Methods and apparatus for characterization of tissue samples
US731054719 Jul 200418 Dic 2007Medispectra, Inc.Fluorescent fiberoptic probe for tissue health discrimination
US73134246 Mar 200325 Dic 2007Critisense Ltd.Diagnosis of body metabolic emergency state
US745969618 Abr 20032 Dic 2008Schomacker Kevin TMethods and apparatus for calibrating spectral data
US746916018 Abr 200323 Dic 2008Banks Perry SMethods and apparatus for evaluating image focus
US74991615 Jul 20063 Mar 2009The Board Of Regents Of The University Of Texas SystemDepth-resolved spectroscopy method and apparatus
US7899518 *12 Sep 20051 Mar 2011Masimo Laboratories, Inc.Non-invasive tissue glucose level monitoring
US800552712 Dic 200723 Ago 2011Luma Imaging CorporationMethod of determining a condition of a tissue
US800828317 Oct 200530 Ago 2011Neurotherapeutics Pharma, Inc.Methods and compositions for the treatment of neuropsychiatric disorders
US812653113 Dic 200128 Feb 2012Boston Scientific Scimed, Inc.Miniature spectrometer
US81401489 Ene 200720 Mar 2012Boston Scientific Scimed Ltd.Readable probe array for in vivo use
US828048430 Abr 20092 Oct 2012The Invention Science Fund I, LlcSystem, devices, and methods for detecting occlusions in a biological subject
US831777619 May 200827 Nov 2012The Invention Science Fund I, LlcCirculatory monitoring systems and methods
US834634618 Ene 20061 Ene 2013The Board Of Trustees Of The Leland Stanford Junior UniversityOptical analysis system and approach therefor
US838026824 Jun 200519 Feb 2013Massachusetts Institute Of TechnologySystem and methods of fluorescence, reflectance and light scattering spectroscopy for measuring tissue characteristics
US840388119 May 200826 Mar 2013The Invention Science Fund I, LlcCirculatory monitoring systems and methods
US840913220 Dic 20072 Abr 2013The Invention Science Fund I, LlcTreatment indications informed by a priori implant information
US863667013 May 200828 Ene 2014The Invention Science Fund I, LlcCirculatory monitoring systems and methods
US866063724 Ene 201225 Feb 2014Boston Scientific Scimed, Inc.Miniature spectrometer
EP1877945A2 *25 Abr 200616 Ene 2008Horiba Jobin Yvon IncMethod for classification of carbon nanotues and other materials
WO1997006724A1 *14 Ago 199627 Feb 1997Maurice AnidjarEndoscopic imaging for the detection of cancer lesions
WO1999013764A1 *11 Sep 199825 Mar 1999Pierre Marc AprahamianDetecting and mapping inflamed zones in a living tissue
WO2001008552A1 *2 Ago 20008 Feb 2001Biophysica LlcSpectroscopic systems and methods for detecting tissue properties
WO2002024048A2 *25 Sep 200128 Mar 2002Avraham MayevskyApparatus and method for monitoring tissue vitality parameters
Clasificaciones
Clasificación de EE.UU.600/476, 600/477, 600/478
Clasificación internacionalA61B5/00
Clasificación cooperativaG01J3/4406, G01N2021/6484, G01N2021/6417, A61B5/0059, G01N21/6456, G01N2021/6419
Clasificación europeaA61B5/00P, G01J3/44B, G01N21/64P4
Eventos legales
FechaCódigoEventoDescripción
15 Sep 2004FPAYFee payment
Year of fee payment: 12
29 Mar 2001SULPSurcharge for late payment
Year of fee payment: 7
29 Mar 2001FPAYFee payment
Year of fee payment: 8
7 Nov 2000REMIMaintenance fee reminder mailed
17 Sep 1996FPAYFee payment
Year of fee payment: 4
8 Abr 1993ASAssignment
Owner name: MASSACHUSETTS INSTITUTE OF TECHNOLOGY, MASSACHUSET
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:BARAGA, JOSEPH;REEL/FRAME:006490/0491
Effective date: 19930323
8 Feb 1993ASAssignment
Owner name: MASSACHUSETTS INSTITUTE OF TECHNOLOGY, MASSACHUSET
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:RICHARDS-KORTUM, REBECCA;REEL/FRAME:006423/0327
Effective date: 19930202
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:RAVA, RICHARD;REEL/FRAME:006419/0087
Effective date: 19930104